Previous work on the development of retinal patterns and nerve circuits indicate that the ommatidial units are formed by cell recruitment (not by a clonal mechanism). Moreover, addition of new elements to the pattern seems to occur "in register with neighboring cells already fixed in a pattern". To test such hypotheses, it seems logical to introduce "a gap" interposed between the differentiated and the undifferentiated cells of the disk. Such gaps will be made by microsurgery. After transplantation and methamorphosis, the resulting heterotopic eyes will be systematically explored (light and TEM) for detecting alterations in the normal arrangement of the photoreceptor cells. Pattern formation in the retina implies some kind of cell communication. Therefore, the fine structure of the intercellular junctions appears as a point of particular interest and the subject will be studied using the Freeze-etch technique.